ECO published report GL 2Sep2009

8/7/2019 ECO published report GL 2Sep2009

http://slidepdf.com/reader/full/eco-published-report-gl-2sep2009 1/36

1

The European Cluster Observatory

EU Cluster Mapping and Strengthening Clusters

in Europe

Center for Strategy and Competitiveness, CSC

Örjan Sölvell

Christian Ketels

Göran Lindqvist



2



3

Background

The concept of clusters originates with Professor Michael Porter, who also has been the

driving force behind cluster mapping methodology. Current cluster mapping efforts in

Europe build on his initial work in the US. Here is an overview of how cluster mapp ing

has evolved and how it led to the current European Cluster Observatory.

In the mid-1980s Professor Michael Porter at the Harvard Business School was

contemplating why some firms – based in particular nations, regions or business

environments – were building globally leading positions while firms in other

environments developed less sophisticated and innovative strategies. It was strikingthat firms in different regions succeeded in different industries and market segments,

even if the regions had simila r levels of prosperity. If firms differed in their ability to

innovate and upgrade, the differences between regions were as striking. As Professor

Porter and other scholars particularly within the field of Economic Geography had

noted earlier, clustering, industrial concentration and regional specialisation were

striking phenomena in all economies. A lso, clusters could be identified in many types of

industries: in high-tech fields and in traditional industries, in handicraft industries and

in manufacturing as well as in services, in small and large firm -dominated industries,

and so on. On a case basis , clusters with a global reach were easily identifiablethroughout a range of industries, including financial services in inner London, film in

Hollywood and Bollywood, watches in Switzerland, flowers in The Netherlands and

Colombia, medical instruments in Massachusetts, and so on. But Professor Porter

decided to develop a general framework for cluster mapping across the U.S. economy,

which led to the U.S. cluster mapping project. This data was then used by several

institutions including the National Governors’ Association and the Council on

Competitiveness in a large number of regional cluster and competitiveness efforts.

In 2000, Professor Porter had worked out his model which could statistically define anddescribe clusters across the U.S. economy. 1 The mapping consisted of two fundamenta l

parts:

• the development of cluster codes which can identify and measure industrial

agglomeration within regions;

1 Clusters of Innovation: Regional Foundations of U.S. Competitiveness . 2001. Washington, DC: Council

of Competitiveness.



4

• the development of performance measures which can measure the

competitiveness and dynamism of clusters.

For the first task, co-location patterns of industries across the U.S. were calculated. Such

industry agglomerations would reflect “revealed patterns of externalities. If two or

more industries tend to co-locate it is a signal that these industries have common

interests or linkages, such as th e sharing of labour skills, technological co -operation and

the like. A set of 41 so -called “traded cluster codes were decided on, accounting for

roughly one third of the total U.S. employment. Cluster performance was measured by

collecting both statistical materials (growth over time, wage data, etc.) and survey data

based on managers’ views. The U.S. model was later adopted by a research team in

Canada.

The establishment of the European Cluster Observatory

ver. 2

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

US SW E ECO ver. 1EU-10

Porter:

• US cluster codes

• Mapping of the US

Lindqvist, Malmberg &Sölvell:

• Translation of US

codes to EU codes

• Mapping of Sweden,

first application in

Europe

Ketels, Sölvell, Lindqvist(IVO):

• Mapping of EU-10

• Star measurement

system developed

Sölvell, Ketels, Lindqvist& Protsiv (CSC):

• Mapping of EU-27+4

• Launch of the

European Cluster

Observatory

Lindqvist, Protsiv ,Sölvell & Ketels (CSC):

• Revision of cluster

codes based on

European data

• Introduction of

prioritised sectors

(e.g. creativeindustries)

In 2003, the U.S. model was brought to Europe by Professor Örjan Sölvell, Dr. Christian

Ketels, and Dr. Göran Lindqvist in Stockholm. A first mapping exercise was done for

Sweden2 and a statistical concordance table was developed in order allow the use U.S.codes on European data .

In 2004, Ivory Tower, a cluster consultancy firm based in Stockholm, was asked as a

subcontractor (Europe INNOVA, EUC -EST, under FP6) to map all clusters of the EU -10

2 Lindqvist, G., Malmberg, A., Sölvell, Ö. 2008. Swedish Cluster Maps: A Statistical Inventory of Clusters

in Sweden. 2002. Stockholm: Center for Strategy and Competitiveness at Stockholm School of Economics.

(Swedish edition from 2003).



5

countries. The project was led by a Panel Group of Experts, under leadership of Mr.

Antoni Subira of Spain. The EU-10 cluster mapping data were published in the first

Europe INNOVA paper. 3 The EU-10 project added a new innovation to clustermapping: the measurement of concentration and specialisation by the use of three

distinct indexes – cluster size, specialisation and regional labour market focus. A few

clusters scored on all three measures and those clusters were designated as “three star ”

clusters. The star methodology was a sound way of describing degree of cluster

agglomeration, and was easily understood by non -experts in the field of clusters, cluster

initiatives and cluster policy.

3 Innovation Clusters in the 10 New Member States of the European Union . 2006. Europe INNOVA paper

No. 1. European Commission Directorate General Enterpri se and Industry.



6

Overview of the European Cluster Mapping Project

Based on the experience with the previous cluster mapping efforts, DG Enterprise and

Industry awarded the contract for the European Clust er Mapping Project to a

consortium coordinated by Professor Sölvell at the Center for Strategy and

Competitiveness (CSC) at the Stockholm School of Economics. It covered all of EU -27

plus Iceland, Israel, Norway, Switzerland, and Turkey. The cluster mappin g part of the

project was later renamed “The European Cluster Observatory (ECO), and the web site

was launched in July 2007. Apart from the cluster mapping effort (driven by CSC), the

project also included a mapping of cluster policies across Europe (cond ucted by a group

of research institutions coordinated by Oxford Research), a number of cluster cases

studies (conducted by the Competitiveness Group in Barcelona), and the coordination

of a high-level advisory group chaired by Senator Pierre Lafitte with s enior

representation from EU member countries (supported by Fondation Sophia Antipolis).

All results of the project and methodology are available at the website of the European

Cluster Observatory, www.clusterobservatory.eu.

The European Cluster Observatory provides unique data on clusters in Europe, e.g.

allowing the user to choose a region (EU/nation/one of 250 regions) and a sector (all orone of 38 categories) and display the corresponding cluster data di rectly on maps of

Europe. In total the European Cluster Observatory lists some 150 three star clusters, 500

two star clusters and just over 1,300 one star clusters (out of a potential of just under

10,000 regional clusters).

The user can also choose to dis play cluster organisations on the map (see figure below).

Already, more than 150 cluster organisations have provided data about themselves to

the Observatory, bringing the current number of cluster organisations represented on

the Observatory to over 1,100 . A Directory of cluster organisations in Europe waslaunched in August 2009.

Many users contact the Observatory to receive more information. Others have added

cluster cases and reports which are continuously uploaded in the Library section.



7

Cluster Mapping

Cluster Organisations

Cluster Policies

Cluster Library

The Observatory’s combination of advanced cluster mapping capabilities with

information about cluster organisations has attracted significant interest. Since its

launch, ECO has tripled its number of users to reach about 4,0 00 unique visitors per

month (June 2009). Over time, a large number of sites around the world dealing with

clusters have posted links to the European Cluster Observatory, leading to more traffic.

The Observatory has also become an important source of information for policy makers.

The European Cluster Memorandum (the text is found in the Appendix) , signed by a

large number of cluster organisations across Europe and presented to the European



8

Commission at a high-level conference in Stockholm in early 2008, was to a large degree

informed by the findings of the Cluster Mapping project. CSC staff has played a central

role in drafting the Memorandum and supporting the discussions of the high -leveladvisory group that presented the Memorandum. Cluster mapping data was supplied

to other EU-sponsored projects, including the BSR InnoNet. CSC has also supported the

Commission staff in the preparation of the Communication 4 on clusters and the Staff

Working Paper5 on clusters and cluster policies presented in late 2008.

In the final month of the project, the CSC team has furthermore launched a new series

of specialised cluster mapping reports, starting with an analysis of knowledge -intensive

business services (KIBS). These reports are intended to present data in ways that are in

line with the organisation of government policies to support decision making in these

areas. CSC has also exploited more granular employment data to test whether the

cluster definitions ultimately created based on agglomeration patterns in the US need to

be further refined given the realities in Europe. While economics play out similarly

across continents, there are areas of economic activity in which Europe has a deeper or

different level of economic organisation that might drive patterns of economic

geography diverging from those observed in the U.S.

Project findings: Understanding the nature and role of clusters

The rich data on clusters generated in this project provides the foundation for

improving the general knowledge about the nature and role of clusters.

Clusters are a part of the microeconomic business environment of a region. Clusters are

shaped by certain almost deterministic forces (blue arrow in figure below) related to the

overall history and culture of a region, the geographical circumstances (access to

waterways, how affluent neighbours are etc.), general institutions and regulations, and

the overall macroeconomic environment. All clusters within a nation are affected bythings such as the exchange rate, colour of the government, and historical and

geographical circumstances of the nation or region.

4 Towards world -class clusters in the European Union. European Commission. Oct 2008.

5 The concept of clusters and cluster policies and their role for competitiveness and innovation: main

statistical results and lessons learned. Commission Staff Workin g Document SEC (2008) 2637.



9

The Funnel Model – The General Business Envi ronment in a Nation

If we take the analysis of the funnel one step further, one must distinguish between

different clusters within a nation or region. Thus, within the same national or regional

context we have a scale of clusters ranging from highly dynamic and competitive onesto more static and uncompetitive ones. In line with this we expect to see more

competitive firms on the right hand side and less competitive firms on the left -hand

side of the scale in the figure below.



10

The Funnel Model: A Range of More or Less Competitive Clusters

Clusters are also shaped bottom -up, from entrepreneurial action and firms

implementing new strategies and business models. Such activities ar e not coordinated

but part of the normal market mechanism – the invisible hand is at work. However, as aresult of these actions, the larger cluster environment will either develop or decline.

Decisions to invest improve and innovate helps to build the clu ster, whereas decisions

to leave the region will push the cluster towards decline, and resources will slowly

merge with other areas of the economy or become idle.



11

The Funnel Model: Entrepreneurship and New Strategies

The evolution of clusters thus emanates from both deterministic (legacy, culture,history) forces and voluntaristic forces. In addition to this we have the constructive,

more conscious forces which will impact the development and competitiveness of the

cluster. One type of constructive force emanates from policy implementing conscious

efforts to improve the microeconomic business environment of a region. Other

constructive forces emanate from initiatives from actors within the cluster, including

civic leaders from private firms, organisations and academia. Local leaders behind

cluster initiatives take on a constructive role to improve the workings of the cluster or

the larger regional environment. Typical objectives of such initiatives includeupgrading of human resources, expansion of the cluster stimulating new firm formation

and attracting new firms to the cluster, business development, and commercial

collaboration such as joint export initiatives or coordinated purchasing to increase

purchasing power. Ot her objectives include upgrading of technology and improving

the overall business environment, including initiating dialogue on new regulations and

upgrading the infrastructure.



12

The Funnel Model: Constructive Forces Shaping the Cluster

Combining the two sets of forces, both evolutionary and constructive, we can gain a

better understanding of how clusters develop , whether they will increase in dynamism

and size, or if they will go into decline.



13

Clusters – Where Evolut ionary and Constructive Forces Meet

We know from our research that cluster dynamics is a highly complex process, and is

best understood as a combination of evolutionary and constructive forces. However,constructors must be aware that the evolutionary forces are strong, and political vision

can easily remain as visions. A large portion of humbleness is in place as constructors

roll up their sleeves.

Clusters and Economic Benefits

The fact that economic activity tends to agglomerate or cluster in particular locations is

driven by efficiency advantages (lowered costs, including transaction costs), flexibility

advantages (high mobility of labour and other resources) and innovation advantages

(knowledge spill-overs and cooperation). The role of clusters in explaining economic

performance of regions has been confirmed in several studies 6. From our own European

data we can show that economic prosperity among the regions of Europe is related to

degree of cluster strength (share of employees in clusters with a location quotient larger

than 2), see figure.

6 Porter, M. E. (2003). The economic performance of regions . Regional Studies, 37(6,7), 549.



14

Regions of Europe: Level of Clustering and Prosperity

y = 83342x 2 - 16467x + 22886

R2 = 0,3941

10 000 €

20 000 €

30 000 €

40 000 €

50 000 €

60 000 €

70 000 €

0% 10% 20% 30% 40% 50% 60% 70% 80%

Source: European Cluster Observatory. ISC/CSC cluster codes 1.0, dataset 20070510

Cluster Strength and Prosperity in Europe

Today, there is plenty of evidence to suggest that innovation and economic growth is

heavily geographically concentrated. Clusters provide an environment conducive to

innovation and knowledge creation. Regions with strong cluster portfolios are

innovative leaders, while regions with no clusters or isolated research facilities fall

behind. Globalisation has increased the benefits of strong clusters and raised the costs

for regions which fail to develop a clear speciali sation profile. Strong clusters emerge i n

open markets where intense rivalry and cooperation within and between clusters

coexist. Clusters emerge, where competition across regions enables companies,

entrepreneurs and financial actors to choose the location of their activities based on theattractiveness of regions, not in response to artificial barriers for cross -border trade and

investment. Globalisation has increased the need to combine strong internal dynamics

within the cluster, with solid linkages to clusters and markets located elsewhere.

Regional specialisation also brings risks, making regions more vulnerable to cluster -

specific demand shocks or fundamental technological shifts. The emerging evidence

suggests, however, that a cluster -based regional economy still generates better



15

outcomes. First, the economic costs of lower productivity due to a lack of speciali sation

have dramatically increased with globally integrated markets. Second, dynamic clusters

that are open to outside trends are better at dealing with external shocks, for exampleby transferring existing skills into new market areas. And third, our research indicates

that the most successful regions tend to have a portfolio of clusters related through

linkages and overlaps that ease the trade -offs between specialisation and diversification.

In Europe, the share of employment in strong clusters tends to be lower than in the

United States. This is particularly visible in medium sized regions. In the US, such

regions tend to be highly specialised with strong presence of employment in a small

number of clusters. In Europe, regions of similar size instead tend to have employment

across a wider range of clusters, with lower levels of true speciali sation in any one ofthem.

Most likely, this is a reflection of past and still remaining barri ers to competition across

regions in Europe. While in the US companies have for a long time been able to choose

their location based on where the production conditions, including the presence of

related and supporting industries, were most beneficial, in E urope the access to the end

market was much more important as a driver of locational decisions. While the

common market programme has made significant progress in reducing these barriers in

Europe, there is clear evidence from companies that markets across Europe remainfragmented, forcing locational decisions to be driven more by market access than

production efficiency considerations (Ketels, 2007) 7.

Clusters and Innovation

Innovation performance tends to be highly skewed across regions, both within nations

and across nations. A large number of empirical studies on regions and innovative

performance have been published in t he last decade (see Crescenzi, Rodríguez -Pose, &Storper, 2007,8 for an excellent overview). The data from the European Cluster

Observatory reveals an important relationship between regional specialisation (degree

of clustering) and innovative performance (measured as patenting levels). Regions in

7 Ketels, Christian (2007). State of the Region Report The Baltic Sea Region as a Place to Do Business, Baltic

Development Forum.

8 Crescenzi, R., Rodríguez -Pose, A., & Storper, M. (2007). The territorial dynamics of innovation: aEurope-United States comparative analysis. Journal of Economic Geography , 7, 673-709.



16

Europe without clusters (i .e. with employment evenly spread out across sectors) are all

performing badly (dots to the left in the figure below). On the other hand, all regions in

Europe with many ranked clusters are all top performers (to the right in figure below).In the group of regions with a few ranked clusters, some are performing well and other

less well. Again, this underlines that economic performance of a region is not only

explained by the degree of speciali sation, but also involves other aspects of the broader

microeconomic business environment, such as labour quality, research and education,

and access to venture capital and advanced infrastructure.

Cluster Strength and Patenting Levels in European Regions

R2

= 0.357

3

4

5

6

7

8

0 0.5 1 1.5 2 2.5 3 3.5 4

Region's cluster strength

Region's

patenting

level

Source: European Cluster Observatory. ISC/CSC cluster codes 1.0, dataset 20070613

Regions with clusters do also perform better in terms of innovation performance. A

competing, but more often complementary, explanation to regional success is the

degree of urbanisation, where metropolises offer diverse and creative environments,



17

and also proximity to academic institutions. Our Research on European data shows that

both urbanisation and regional specialisation, i.e. clustering, bring economic prosperity

to regions, but in different ways. Urbanisation has a direct effect on regionalperformance, whereas clusters work through the process of being more innovative

environments, which in turn leads to economic prosperity 9.

9 Lindqvist, Göran, Sergiy Protsiv, Sölvell, Örjan. Regions, Innovation and Economic Prosperity: Evidence from

Europe, mimeo. Stockholm School of Economics, Stockholm: 2008.



18

Project findings: Clusters as a tool for policyThe project has also made a significant contribution to the understanding of the cur rent

state and the logic of using clusters as an instrument of economic policy.

Economists consider policy interventions as justified when specific conditions exist that

reduce the ability of the normal market process to lead to optimal outcomes from an

overall welfare perspective. Such ”market failures” provide the traditional motivation

for economic policy. The local externalities that give rise to clusters create a number of

such market failures:

• Coordination failures exist, because individual companies consider in their

decisions, be it whether to locate in a cluster or what investments to undertake

being there, only the impact on themselves, not on others.

• Information asymmetries exist, because even if the incentive problems of taking

account of the impact of own actions on others could be managed, the knowledge

necessary to make the right ”social” decision is dispersed among the many

participants of the cluster.

• Path dependency exists, because decisions not only influence the present, but also

the possible evolutionary path of the cluster in the future. Both coordination

failures and information asymmetries thus have a dynamic dimension as well.

And social and private discount rates might differ, creating an additional source of

market failure.

Where cluster policy addresses market failures, it does not reduce global welfare. Under

some assumptions, the free competition between rational governments in supporting

clusters even leads to the best possible outcome, not a race to the bottom. 10 While thesearguments do not prescribe specific policy interventions, they give some guidance on

the direction that cluster policy should take. The best approach is always to target the

market failure at the source. Policy can subsidi se activities that are underprovided

because of coordination failures or differences in discount factors. And policy can

10 Norman, Victor, Anthony Venables, Industrial Clusters: Equilibrium, Welfare and Policy, Economica,

Vol. 71 (2004), 543 – 558



19

facilitate platforms for collective action to overcome coordination failures and

informational asymmetries. In practice, efforts to address market failure are never

perfect.11 They suffer from government failure in implementation (lack of knowledge totarget the intervention, inability to provide incentive -neutral financing, political

pressure by interest groups for beneficial treatment, etc.) and might have unintended

side-effects, creating collateral costs that outweigh the benefits.

Economic policies can be compared on both the impact that they generate, i.e.

addressing the problem or market failure, and the costs they might create, i.e.

distortions or government failure. Policies that target individual companies are highly

effective but also very distortionary. Policies that target the entire economy have little if

any distortionary effect but are often also not very effective. Policies targeted atindividual industries c ome somewhere in the middle on both accounts. Cluster policy,

however, offers a superior mix of benefits and costs. It is organi sed around a group of

industries that by definition have strong linkages. Targeting policy at them will thus not

only be effective but even trigger additional benefits from positive spill -overs that are

induced. And while the policy is neutral within the cluster where competition for

factors of production is the strongest, it is distortionary only relative to activities outside

the cluster where by definition other skills and assets are needed. While some distortion

remains, the approach promises a potentially better balance of effects.

Two opposing approaches to cluster policy

In the academic debate, the strongest criticism of cluster policy does not come from

researchers that claim that locational factors are irrelevant, but from economic

geographers and others that fully support the view that locational fact ors are important.

Some criticise the “fuzzy” way the cluster framework is translated from an academic

idea into a practical policy concept. 12 But while the issues raised in these discussions

reflect important operational challenges of implementing cluster policy, they also tendto reveal a limited sense of the needs of policy practitioners . Cluster policy is a complex

process and requires a framework that enables context -dependent on-the-ground

11 Rodrik, Dani, Normalizing Industrial Policy , Working Paper No. 3, Commission on Growth and

Development, Washington, D.C.: 200 8.

12 Martin, R, Peter Sunley, Deconstructing clusters: chaotic concept or policy panacea? Journal of Economic

Geography, Vol.3 (2003), No. 1, 5 -35.



20

choices, but this does not provide a conceptual argument against its use. Others provide

a more fundamental criticism of the motivation for cluster policy 13 that turns out to be

highly revealing for how the lack of a generally acce pted definition of cluster policycontinues to hamper the debate.

To understand the different views on cluster policy, it is useful to go back to a simple

diagram that relates agglomeration to competitiveness. The evidence discussed in

chapter 2 points towards a positive relationship between the two, a fact that is generally

accepted by critics as well as supporters of cluster policy (as discussed previously there

are differences in the view on how strong this relationship is relative to other factors).

But how should cluster policy intervene to move a location from a place at the bottom

left to the top right? This is where the fundamental difference sets in.

One approach sees agglomeration as the central policy lever; as agglomeration rises,

competitiveness will naturally follow as cluster effects set in. With agglomeration the

ultimate goal, efforts to attract companies through incentives from tax rebates to free

infrastructure naturally come to the forefront of the policy debate. Economic

geography-based approaches, too, look at the effects of traditional tax, trade, and

regional policies on agglomeration patterns. 14 Dynamic ”new economic geography”

models provide guidance on when and how these instruments should be used to have a

maximum impact:15 the process of agglomeration in these models is characteri sed by

important break-points at which economic geography patterns are determined. For

economic policy, this implies that intervention has to be early, i.e. at a time when the

locational patterns of where a dominant cluster will be located has not been determined

yet. And it has to be massive, i.e. it has to give such a meaningful boost that the location

gains sufficient critical mass to be far ahead of all potential rivals. And it im plies a

critical role for identifying a small number of clusters on which economic development

then hinges.

13 Duranton, Gilles, Henry Overman, Exploring the Detailed Locational Patterns of UK ManufacturingIndustries using Microgeographic Data, Journal of Regional Science , Vol. 48 (2008), No. 1, 213 – 243

14 Baldwin, Richard, Rikard Forslid, Philippe Martin, Gianmarco Ottaviano, Frederic Robert -Nicoud,Economic Geography and Public Policy , Princeton University Press, Princeton: 2003

15 Brenner, Thomas, Cluster Dynamics and Policy Implications, Zeitschrift für Wirtsc haftsgeographie, Vol.

52 (2008), 146-162. And Brenner, Thomas, Policy Measures to Create Localised Industrial Clusters, in T.

Brenner und D. Fornahl (Hrsg.): Cooperation, Networks, and Institutions on Regional Innovation

Systems, Cheltenham: Edward Elgar, 2003, 325-349.



21

If large-scale targeted subsidies in the early phase of cluster emergence are the policies

under discussion, should they be used? Not only critics of cluster policy come to a

negative answer: such policies are likely to fail because they require an abundance ofinformation and ability in the hands of the policy maker. And they are not even

necessary: current economic geography is already in line with the fundamentals

including local externalities, so any policies to change the location of companies would

lead away from an existing optimum. 16

Two Perspectives on Cluster

Development

MORE(Agglomeration)

BETTER(Competitiveness)

FINISH

Another approach sees competitiveness as the central policy lever; as competitiveness

rises, agglomeration will naturally increase as the cluste r becomes more attractive for

new entrants.17 With competitiveness as the ultimate goal, clusters become a process

tool to design and implement policies more effectively, not an ultimate objective. The

instruments then targeted at existing clusters are well known from innovation policy,

regional policy, and enterprise policy. They are supplemented by actions that

specifically support collaboration in their use and that create platforms for collaborationwithin an agglomeration. The competitiveness literature, including the insights on

cluster evolution provide guidance on when and how to use these instruments that is

16 Martin, Philippe, Thierry Mayer, Florian Mayneris, Natural clusters: Why policies promoting agglomeration

are unnecessary, VOX: 4 July 2008

17 Roriguez-Clare, Andres, Coordination failures, clusters and microeconomic interventions , IADB Working

Paper 544, Washington: December 2005a



22

radically different from the model cluster policy critics have in mind . The focus should

be largely on agglomerations that have already passed the test of the early stages of

development.18 This indicates that the fundamental conditions for economic success arepresent and active collaboration can b ecome a ”turbo” for the use of strengths already

in place. The focus of policy interventions should be on enabling collaboration and

channelling existing resources in a different way, using moderate amounts of new

funding. Large new funds are not necessary and could be harmful by increasing the

potential for distorting incentives. And while a selection of clusters is necessary to be

able to deploy sufficient resources and attention on any one initiative, economic

development is the result of many clusters i n all regions flourishing, not just a few per

country.

If these are the policies under discussion, should they be used? Even the critics of

cluster policy have a slightly favourable view . Improvements in the fundamentals of

competitiveness are a sensible goal and the suggested approach limits the downside.

But they remain skeptical about whether cluster efforts can have a sufficiently strong

impact on improving underlying competitiveness. The quantitative evidence is still

limited but points to moderate po sitive effects.19 Proponents of cluster policy see

enough case-evidence that such efforts can in fact lead to a much more meaningful

improvement in the way policies for higher competitiveness are being conducted. 20

18 Roriguez-Clare, Andres, Clusters and Comparative Advantage: Implications for industrial policy ,

Pennsylvania State University, mimeo. , University Park: September 2005b

19 Engel, Dirk, Oliver Heneric, Stimuliert der BioRegio-Wettbewerb die Bildung von Biotechnoligieclustern in

Deutschland?, ZEW Discussion Paper 05 -54, Mannheim: 2004.

Dohse, Dirk, Cluster -based Technology Policy: The German Experience, Industry and Innovation , Vol. 14,

No. 1 (2007), pp. 69 – 94.

Dohse, Dirk, Tanja Staehler, BioRegio, BioProfile and the Rise of the German Biotech Industry , Kiel InstituteWorking Paper, No. 1456, Kiel: October 2008.

Falck, Oliver, Stephan Heblich, Stefan Kipar, The Extension of Clusters: Differences -in-Difference Evidence

from the Bavarian State-Wide Cluster Policy , Jena Economic Research Papers, 2008 -073, Jena: 2008.

Fromholt-Eisebith, Martina, Günter Eisebith, Looking Behind Facades: Evaluating Effects o f (Automotive)

Cluster Promotion, Regional Studies , 2008.

20 Waits, May Jo, The Added Value of the Industry Cluster Approach to Economic Analysis, StrategyDevelopment, and Service Delivery, Economic Development Quarterly , Vol. 14 (2000), 35-50.

Cortright, Joseph, Making Sense of Clusters: Regional Competitiveness and Economic Development , Brookings

Institute, Washington, D.C.: March 2006.



23

There remains a fair amount of disagreement in the debate about cluster policies. At

least part of this disagreement is related to a lack of effective communication between

theoretical research and policy practice. This communication failure leads to afundamental disconnect on what cluster policy is and how it is related to

competitiveness upgrading. For many researchers, improving competitiveness is

fundamentally an automatic process, driven by the self -interest of all parties involv ed.

For most practitioners, improving competitiveness is a complex challenge of identifying

action priorities and mobilising allies to implement them. Cluster policy, as understood

by its proponents, is an answer to these real challenges that practitioners face,

challenges that the critics assume will being taken care of automatically over time.

Implementing cluster policy to improve competitiveness

The discussion so far has established a solid conceptual argument for cluster policy as a

tool to leverage cluster agglomerations as a tool to achieve higher impact on upgrading

underlying competitiveness. Whether these possible benefits of such policies m aterialise

in a meaningful fashion, is a question of how and where they are implemented, not just

of their conceptual approach. Three issues are of particular important. First, does cluster

policy open the door to distortive interventions that have little to do with the original

objective but easily follow once cluster program mes are launched? Second, are theeffects of cluster policy strong enough to warrant more fundamental policy interest?

Third, which locations should use cluster policy?

Cluster policy uses industry-specific policy instruments and activities. As such, it can

become a politically convenient cover for what then in reality is nothing else but

traditional distortive industrial policy. The political economy argument that some

critics then make is the following. Even if cluster policy has its merits, it opens the

political process for all kinds of sector -specific interventions that undo its theoretical

benefits. 21 On balance, they argue, it is then better to forgo a useful instrument likecluster policy if it leads to opening the Pandora box of ”vertical” policies.22 This is an

Mills, Karen, Elizabeth Reynolds, Andrew Reamer, Clusters and Competitiveness: A New Federal Role for

Stimulating Regional Economies , Metropolitan Policy Programme , Brookings, Washington, D.C.: 2008.

21 Rodrik, Dani, Normalizing Industrial Policy , Working Paper No. 3, Commission on Growth and

Development, Washington, D.C.: 2008.

22 EBRD, Transition Report 2008: Growth in Transition , European Bank for Reconstruction and

Development, London: November 2008.



24

important consideration. But it has to be balanced against another political economy

dynamic. Many governments are under intense political pressure to ”do more” rather

than upgrading the general business environment. In such situations, the alternative tocluster policies is often not the absence of targeted policy action, but the use of exactly

the type of old style industrial policy tools that should be avoided. And spe cific steps

and conditions can reduce the likelihood of cluster policies being high -jacked by narrow

interest groups: High exposure to external competition and robust competition policies

domestically reduces the danger that collaboration leads to lower ra ther than more

sophisticated rivalry. Competition models with the involvement of external jurors can

de-politicise the selection process and induce a clear orientation to excellence. And the

threat of losing funding in case cluster dynamics remain low avoi ds subsidising many

weak clusters rather than allowing stronger clusters to gain position. Overall, especially

the role of government needs to be carefully designed. While there is no systematic

evidence that a government role per se is negative for Cluster Initiativ es,23 government

cannot create clusters 24 and can easily impose conditions that hurt competitiveness.

Cluster policy has in the past often been applied at the level of individual clusters. But

simple arithmetic suggests that working with one cluster in a region, even if it is a large

one, is unlikely to generate economic outcomes that are meaningful for an overall

regional economy. The average regional cluster accounts for about 1% of totalemployment in a region (European Cluster Observatory, 2008); larger cluster categ ories

like financial services or transportation can in individual cases reach much higher levels

but are for most regions not above 5% of total regional employment. High -tech clusters

like biotech range at a fraction of such numbers. Purely growing one suc h cluster by

improving its competitiveness can thus have high impact on a few individuals and

companies but will tend to have only a moderate impact on the regional economy at

large. A number of recent analyses have identified how cluster policy can be des igned

to affect the wider regional economy and thus become a quantitatively important tool

for economic development efforts. 25 Locations should take a portfolio perspective on

23

Sölvell, Örjan , Göran Lindqvist, Christian Ketels, The Cluster Initiative Greenbook , Ivory Tower,

Stockholm: 2003.

24 Porter, Michael E., quoted in “Politiker kan inte skapa kluster, hävdar klustrens egen pappa, Svenska

Dagbladet , 28/1/2008.

25 Pietrobelli, Carlo, Roberta Rabelotti, Upgrading in Clusters and Value Chains in Latin America: The Role of

Policies, Inter-American Development Bank, Washington, D.C.: January 2004.



25

their cluster efforts, addressing the different needs of clusters at different s tages of

development and leveraging the linkages across clusters. Effective cluster policy

mobilises all clusters, not just one that is supposed to drive future economic growth.Locations should leverage the experience of the cluster efforts for economy -wide

improvements. At least part of the business environment weaknesses that create

problems for specific clusters usually also affect companies more generally. And

locations should integrate their cluster efforts into a broader economic strategy that

identifies the specific value that the location provides relative to its peers. Clusters can

effectively communicate the unique advantages locations offer, much better than

general attributes like “open for business” or “ entrepreneurial”.

High Level Advisory Group on Clusters, The European Cluster Memorandum , Stockholm: January 2008.

Ketels, Christian, Örjan Sölvell, Innovation Clusters in the 10 New Memb er States of the European Union ,

Europe INNOVA Paper No. 1, European Commission - DG Enterprise and Industry, Brussels, 2006.



26

Implications for future efforts and policyBased on the project findings, a number of conclusions can be drawn for both the next

stage of cluster mapping and the further development of cluster policy.

Cluster mapping

The European cluster mapping data has for the first time off ered a consistent view of

the economic geography of clusters in Europe. This has been an important step forward

for cluster research and cluster policy. However, the discussions with users haveindicated data quality, especially granularity and available i ndicators, are below the

level desired by practitioners and policy makers. A number of issues stand out:

• Practitioners need more data on their clusters and better fit of definitions to their

individual clusters.

• Policy makers get general support for the n otion that clusters matter, but not

enough data on the impact of cluster policy

• Policy makers need more data that can be translated into action recommendationson what to do and what areas to focus on

Some of these issues can be addressed by further data collection in the context of the

Cluster Observatory. But there are also clear limitations that can only be addressed

through the slow process of improving the way statistical data is collected or an entirely

different approach to data collection. Overall, cluster mapping has to be transformed

from a tool that is useful to researchers and high -level policy advice to an instrument

with direct applicability to cluster practitioners and cluster -programme implementing

agencies.

Cluster initiatives

While the cluster mapping project has focused largely on the collection of statistical

data on the agglomeration of economic activities in related fields, the overall effort as

also created additional insights into the needs of cluster initiatives. Among the issues

that came up are the following:



27

• There is clear need for professionali sing the management of cluster initiatives ,

from spreading of general best practices in the field to the use of cluster data by

cluster initiatives. In recent years there has b een an explosion in the demand andsupply of training courses and educational programmes in this area.

Consolidation and standard setting are the usual processes that tend to set in as

professional fields reach some level of maturity. Professional organi sations like

TCI play an important role in such a process, but organi sations like the EU and

other government bodies financing cluster efforts are clearly important as well.

• Impact measurement is a theme that is high on the agenda for many cluster

initiatives as well as for public bodies that manage cluster policy program mes.

Some project evaluations have already been done (for a discussion see Sölvell2009; Chapter 6), but while these efforts have created useful advice on how to

improve cluster programmes, there has been limited if any success in quantifying

the economic impact and return of cluster policies. While cluster mapping might

play some role in this respect in the future, the discussion of cluster policy as a tool

to improve competitiveness rather t han as a direct way to change agglomeration

patterns casts some doubts on relying on this instrument alone. There have been

efforts in the past like the Foundation Clusters and Competitiveness but they have

not reached momentum. Other bottom -up efforts are underway in TCI and as an

initiative of individual regions like Värmland in Sweden. But there is a clear need

for top-down structures as well to allow these initiatives to reach their full

potential.

• One area that has been pushed much higher on the agenda of cluster initiatives in

the last few years is cluster internationalisation . While companies were heavily

engaged in globalising their efforts, cluster initiatives were mainly seen as a tool to

strengthen the local networks and capabilities. But cluster initiatives can also

collaborate internationally, providing the platform for easier internationali sationof companies with other regions that have similar economic profiles. And the

interaction of clusters, facilitated by cluster initiatives, can play an i mportant role

in the creation of global value chains of clusters. To date, several initiatives have

had the aim of building such collaboration networks, but each of them has had a

limited scope. For example, the INTERREG IIIC Project CLOE (Clusters Linked

Over Europe) managed to link only some dozens of cluster organisations,

irrespective of their sectoral affiliation. In the INTERREG IIIC project INNOFIRE,

a web-based co-operation platform was developed for companies, but only in the



28

medical technology sector. In the Europe INNOVA TCAS project there was

another single-sector database for companies developed focusing on the

automotive sector. Under INTERREG IVC one project (Clusters and Cities,CLUSNET) is engaged in trans -national cluster collaboration. The re are also many

publically available databases on cluster organisations. For example, the

association TCI and Europa InterCluster lists several cluster organisations. On the

internet platform XING there is a “cluster -group with persons interested in clus ter

issues. Such fragmented attempts are numerous, but Europe still lacks a full -

fledged, efficient and low-cost information system incorporating data on all

sectors and regions that enables a huge visibility of the cluster organisations – and

their members, especially innovative SMEs. The EU again has a clear role in

removing barriers, for example in the funding rules, and encouraging the

collaboration of cluster initiatives.

Cluster policy

The project has also generated a number of important insights into cluster policy.

Among the key observations are the following:

• Cluster-policies versus framework policies ; the conceptual work on cluster

emergence has pointed out the dual role of creating and emerging forces as drivers

of cluster evolution. The look at t he interplay of existing cluster policies and

existing framework conditions have highlighted their different and

complimentary roles. In the US, cluster policy has been relatively weak but the

strengths of the fundamental framework conditions, especially t he open

competition across the integrated US market, enabled the emergence of specialised

regions with strong cluster structures. Private -sector-led cluster efforts then

compensated somewhat for the lack of coordinated government policy in

mobilising these clusters. But the framework conditions are key to understand ingwhy the US has a stronger cluster structure than Europe. In Europe, cluster policy

has recently become much stronger and has significant potential to mobilise the

inherent capabilities of many European clusters. But the experience in the US

indicates that it is very important not to view these cluster policies as a substitute

for further market integration. In fact, cluster policies will have much more

potential if the framework conditions in Europe will also support the emergence of

a stronger cluster landscape. For the EU this implies that the common market



29

remains a critical element of any cluster -oriented policy. Clusters emerge not as a

consequence of intervention or the financing of clust er initiatives; they emerge in

reaction to fundamentals. This is one of the key lessons that is starting to emergefrom the analysis of the evidence on clusters and cluster policies. The EU plays a

crucial role in establishing this link and designing its p olicies accordingly.

• Financing cluster initiatives versus organi sing policies around clusters . Most of

the current cluster efforts have instinctively focused on supporting cluster

initiatives, i.e. the networks of actors with an agglomeration of related ec onomic

activities. While this is an important way to address the market failures at the

heart of clusters, it is not the only one. There is an important opportunity to

organise existing government policies around clusters, in areas like innovation,workforce development, FDI attraction, SME development, and others. This could

improve the effectiveness of these policies and create the impact governments

wish to reach, while avoiding the market distortions that interventions on behalf

of individual companies or industries tend to create. This way clusters would

become an organising principle to integrate different economic policies from the

perspective of the user, overcoming the policy silos that currently often

characterise economic policy. Cluster policy w ould then be not another such silo,

but a way to overcome the existing separation of efforts.

The EU has a crucial role to play in this respect. Different parts of the EU

Commission target clusters with a multitude of program mes. There is huge

potential to increase the effectiveness of these efforts by using clusters as an

integrating mechanism. The EU has the potential to develop new structures

through which different functional policy program mes can be integrated and then

be made accessible more efficien tly to clusters.

• Impact of cluster efforts . Many evaluations of cluster initiatives suggest that the

individual projects have been successful, but often only for the partners directlyinvolved while the overall impact on the regional or even national econo my has

been negligible. This would suggest the cluster initiatives are a useful tool, but

ultimately not much more than one of many technical instruments at the disposal

of policy makers (framework policies supporting the natural emergence of strong

clusters as in the US might still be much more powerful). But the discussions in

the context of this project and beyond indicate that there are ways in which cluster



30

policy can become much more ”scalable” in its impact.26

One important way to achieve this is the introduction of strong feedback

mechanisms to cross -cutting policies. Cluster initi atives can becoming testinggrounds for the identification of problems and solutions that are then rolled out

more broadly, creating benefits for all companies, not only those that participate in

the cluster effort. So far, few cluster policies are designe d to encourage such

feedback mechanisms. The EU could take the lead in integrating such mechanisms

in efforts it supports.

Another important approach is to put individual cluster efforts in a strong

locational perspective, i.e. see them as part of the over all cluster portfolio of a

region or nation. Governments can then make sure that they have appropriate

policies for their existing strong clusters, their emerging clusters, and general

conditions that enable entirely new efforts to emerge. All of them requ ire different

types of policies but together this can be a very powerful way to organi se policy. It

can overcome the low effectiveness of the cluster -blind policies of the past. But it

can also combine the approach of isolated cluster efforts without integ ration into a

broader regional or national perspective that is characteristic even of many of the

best current cluster programmes. The EU has a critical role to play in providing

the data and tools to European regions that are willing to pursue such an

approach.

26 Ketels, Christian, Clusters, Cluster Policy, and Swedish Competitiveness , Expert Report No. 30, Swedish

Globalisation Council, Stockholm: March 2009.



31

Next stepsThe current Observatory represents the state of the art in providing a Europe -wide

service for cluster organisations. But it is clear that much more is needed for the

Observatory to become a full -fledged information service. To begin with , the cluster

mapping service needs to provide richer data in a way that is more relevant for users of

all types. But above all, it needs to provide a completely new set of services for cluster

organisations and their members supporting collaboration betwe en cluster

organisations and their members.

The proposed next phase of the European Cluster Observatory (ECO -II, 2009-2012)

includes several components. First, the most fundamental change is that ECO moves

from a web based database to a real tool for collaboration (benchmarking, finding

partners in trans-national networks, etc.) that can be used by SMEs and managers of

cluster organisations. Second, the database will be enhanced with new data including

new cluster codes, new employment data, n ew performance data, and new data on

regional business framework conditions. Third, the new Observatory will move from a

rigid functionality to a dynamic user-friendly functionality where the user can

customise their own regions (e.g. the Øresund region cu tting across Denmark andSweden), and customise their own cluster categories (e.g. combining textile, apparel

and footwear into “fashion). See the figure below.



32

Phase I Phase II

Fourth, ECO-II will be designed to be user-driven, allowing for SMEs and cluster

organisations to upload content themselves. Fifth, a new web appearance will enhance

user-friendliness. All these improvements will address identified weaknesses with the

Observatory in its current form.

We have identified four major weaknesses, related to data availability and cluster

mapping, where the Observatory is not delivering enough value today.

• Capturing emerging industries where statistical codes are not relevant: to solve

this ECO-II will conduct qualitative studies, for example focusing on identified

Lead markets .

• Covering relevant and politically prioritised sectors: to solve this ECO -II proposes

to revise cluster codes (based on European data, moving away from the U.S.

codes developed by Professor Porter), and to o ffer custom cluster categories bothdefined by the Observatory (mega -clusters and identified prioritised sectors), and

also allowing for users to develop their own categories.

• Covering relevant regions such as cross -national regions: to solve this ECO -II will

allow for users to identify their own regions, and will also produce data and

analysis for trans-national regional initiatives in Europe.



33

• Telling a richer story about cluster growth and competitiveness: to solve this ECO -

II will develop rich data on regional microeconomic business conditions

(infrastructure, skills, R&D, innovation, entrepreneurship, international clusterlinkages, demand sophistication etc.), making use of secondary data (e.g. CIS and

other sources) complemented with regional survey da ta. In addition cluster

performance data will be added (e.g. based on wages from LFS).

New design,

user-fr iendly interface

Collaboration platform

Cluster mapping andbusiness environment data

Dynamic documentrepository

Cluster mapping

Custom mapping tool

Phase I Phase II

Static mapping tool

Static documents

Cluster organisation data

Trans-national clusternetworks

Cross-nation regions

Lead markets

The new Observatory will build close ties with the European Cluster Alliance (PRO

INNO Strand 2), the Cluster Academy (PRO INNO Strand 6), the Eco-Observatory

(Europe INNOVA Strand 5) and other EU initiatives.



34

An advisory group around Cluster Management will be set up. This advisory group

will consist of local/regional organisations involved in cluster program mes, with

particular interest in cluster programme evaluation.



35

Reports from CSC staff in conjunction with the project

Ketels, Christian , Clusters, Cluster Policy, and Swedish Competitiveness , Expert Report No.

30, Swedish Globalisation Council, Stockholm: March 2009.

Ketels, Christian, Competitiveness in Developing Economies: The Role of Clusters and

Cross-Cutting Policies, In: Nurturing the Sources of Growth in Tanzania -- Workshop

Proceedings , Tanzanian Ministry of Planning, Economy, and Empowerment, Dar -es-

Salam: 2006

Ketels, Christian, Göran Lindqvist, Örjan Sölvell, Cluster Initiatives in Developing andTransition Economies , Center for Strategy and Competitiveness (SSE), Stockholm: May

2006.

Ketels, Christian, Olga Memedovic, From Clusters to Cluster-Based Economic

Development, International Journal of Technological Learning, Innovation, and Development ,

Vol. 1 (2008), 375 – 392.

Ketels, Christian, Örjan Sölvell, Innovation Clusters in the 10 New Member States of the

European Union, Europe INNOVA Paper No. 1, European Commission - DG Enterpriseand Industry, Brussels, 2006.

Lindqvist, Göran, Disentangling Clusters: Agglomeration and Proximity Effects . Published

Doctoral Dissertation. Center for Strategy and Competitiveness, CSC, Stockholm. 2009.

Lindqvist, Göran, Sergiy Protsiv, Sölvell, Örjan, Regions, Innovation and Economic

Prosperity: Evidence from Europe , mimeo. Stockholm School of Economics, Stockholm:

2008.

Sölvell, Örjan, Clusters: Balancing Evolutionary and Constructive Forces , Ivory Tower,Stockholm: 2009.

Örjan Sölvell Ketels, Christian, Göran Lindqvist, Industrial specialization and regional

clusters in the ten new EU member states. Competitiveness Review. Vol. 18, No. 1/2. pp.

104-130. 2008.



36

Other Reports from the Project

High Level Advisory Group on Clusters, The European Cluster Memorandum , Stockholm:January 2008.

Competitiveness Group. Case studies of clustering efforts in Europe: Analysis of thei r

potential for promoting innovation and competitiveness. 2008.

Oxford Research, Cluster Policy in Europe , Europe INNOVA Cluster Mapping Project,

Oslo: January 2008.

Priority Sector Report: Knowledge Intensive Business Services (KIBS). European Cluster

Observatory and Center for Strategy and Competitiveness, Stockholm. 2009.

ECO published report GL 2Sep2009

Documents